1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
3 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=CHECK,SSE,SSE41
4 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE42
5 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
6 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
7 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX,AVX512,AVX512F
8 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX,AVX512,AVX512BW
10 declare i32 @llvm.usub.sat.i32 (i32, i32)
11 declare i64 @llvm.usub.sat.i64 (i64, i64)
12 declare <8 x i16> @llvm.usub.sat.v8i16(<8 x i16>, <8 x i16>)
14 ; fold (usub_sat x, undef) -> 0
15 define i32 @combine_undef_i32(i32 %a0) {
16 ; CHECK-LABEL: combine_undef_i32:
18 ; CHECK-NEXT: xorl %eax, %eax
20 %res = call i32 @llvm.usub.sat.i32(i32 %a0, i32 undef)
24 define <8 x i16> @combine_undef_v8i16(<8 x i16> %a0) {
25 ; SSE-LABEL: combine_undef_v8i16:
27 ; SSE-NEXT: xorps %xmm0, %xmm0
30 ; AVX-LABEL: combine_undef_v8i16:
32 ; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
34 %res = call <8 x i16> @llvm.usub.sat.v8i16(<8 x i16> undef, <8 x i16> %a0)
38 ; fold (usub_sat c1, c2) -> c3
39 define i32 @combine_constfold_i32() {
40 ; CHECK-LABEL: combine_constfold_i32:
42 ; CHECK-NEXT: xorl %eax, %eax
44 %res = call i32 @llvm.usub.sat.i32(i32 100, i32 4294967295)
48 define <8 x i16> @combine_constfold_v8i16() {
49 ; SSE-LABEL: combine_constfold_v8i16:
51 ; SSE-NEXT: movaps {{.*#+}} xmm0 = [0,0,254,0,65534,0,0,0]
54 ; AVX-LABEL: combine_constfold_v8i16:
56 ; AVX-NEXT: vmovaps {{.*#+}} xmm0 = [0,0,254,0,65534,0,0,0]
58 %res = call <8 x i16> @llvm.usub.sat.v8i16(<8 x i16> <i16 0, i16 1, i16 255, i16 65535, i16 -1, i16 -255, i16 -65535, i16 1>, <8 x i16> <i16 1, i16 65535, i16 1, i16 65535, i16 1, i16 65535, i16 1, i16 65535>)
62 define <8 x i16> @combine_constfold_undef_v8i16() {
63 ; SSE-LABEL: combine_constfold_undef_v8i16:
65 ; SSE-NEXT: movaps {{.*#+}} xmm0 = [0,0,0,0,65534,0,0,0]
68 ; AVX-LABEL: combine_constfold_undef_v8i16:
70 ; AVX-NEXT: vmovaps {{.*#+}} xmm0 = [0,0,0,0,65534,0,0,0]
72 %res = call <8 x i16> @llvm.usub.sat.v8i16(<8 x i16> <i16 undef, i16 1, i16 undef, i16 65535, i16 -1, i16 -255, i16 -65535, i16 1>, <8 x i16> <i16 1, i16 undef, i16 undef, i16 65535, i16 1, i16 65535, i16 1, i16 65535>)
76 ; fold (usub_sat x, 0) -> x
77 define i32 @combine_zero_i32(i32 %a0) {
78 ; CHECK-LABEL: combine_zero_i32:
80 ; CHECK-NEXT: movl %edi, %eax
82 %1 = call i32 @llvm.usub.sat.i32(i32 %a0, i32 0)
86 define <8 x i16> @combine_zero_v8i16(<8 x i16> %a0) {
87 ; CHECK-LABEL: combine_zero_v8i16:
90 %1 = call <8 x i16> @llvm.usub.sat.v8i16(<8 x i16> %a0, <8 x i16> zeroinitializer)
94 ; fold (usub_sat x, x) -> 0
95 define i32 @combine_self_i32(i32 %a0) {
96 ; CHECK-LABEL: combine_self_i32:
98 ; CHECK-NEXT: xorl %eax, %eax
100 %1 = call i32 @llvm.usub.sat.i32(i32 %a0, i32 %a0)
104 define <8 x i16> @combine_self_v8i16(<8 x i16> %a0) {
105 ; SSE-LABEL: combine_self_v8i16:
107 ; SSE-NEXT: xorps %xmm0, %xmm0
110 ; AVX-LABEL: combine_self_v8i16:
112 ; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
114 %1 = call <8 x i16> @llvm.usub.sat.v8i16(<8 x i16> %a0, <8 x i16> %a0)